关键词: 立方体模型, 扩展卡尔曼滤波（EKF）, 运动汽车, 跟踪算法

Abstract: This paper presents a real-time monitoring of moving vehicles detection and tracking algorithm under wide range complex transportation scenes with a fixed camera.According to the rigid body characteristic of vehicles，uses the cube model matching with the moving cars image，avoids the three dimensional reconstruction problem，and obtains the physique information of moving cars intuitively.Combined with the non-linear motion characteristic of vehicles，the paper introduces the EKF algorithm，it can approximately forecast the related parameter information of the target，and reduce the search time of the target tracking，therefore it can track the object with real-time and accurately.Through the real transportation video simulation experiment，the algorithm has a good tracking ability and anti-disturbing capability.

Key words: cube model, Extend Kalman Filter（EKF）, moving car, tracking algorithm